I currently use prepackaged Perkin-Elmer/Applied Biosystems Division (PE/ABD) reagents and solvents for the reasons described above, but alternative vendors for solvents and reagents for Applied Biosystems 470A and 477A protein sequencers that I have used in the past include: Burdick and Jackson's acetonitrile (solvent B, S4, and R5), ethylacetate (S2) and n-butylchloride (S1); Aldrich's tetrahydrofuran without stabilizers; and a Millipore Milli-Q system for the water component of solvent A and S4. Other vendors mentioned in the mailing list correspondence included: HPLC grade water from Fisher, EM Science, and Pierce; ethylacetate from Aldrich and Pierce; n-butylchloride from EM Science and Pierce; and Fisher Optima acetonitrile. Alternate reagent vendors were recommended by the e-mail correspondents based on anecdotal experience-comparative data showing that sequencing results were not affected by use of homemade reagents were not provided.
There are many things to bear in mind when making homemade sequencing reagents. Some prepackaged reagents may contain additional chemicals not listed on the label, such as dithiothreitol or other stabilizers. The presence of contaminants and particles from glassware, filters, and dust can result in costly repairs. Reused reagent bottles can form tiny cracks at the rim, leading to particulate contamination and to insufficient delivery volumes. One common problem affecting PTH-amino acid (PTH-AA) identification that is more prevalent with homemade than with prepackaged solvents is the presence of peroxides in solvent A. For PE/ABD and Beckman/Porton instruments, the addition of tetrahydrofuran to solvent A or improper storage of solvent A can introduce peroxides affecting PTH-AA yields, especially oxidation-sensitive amino acids such as lysine and methionine. Tetrahydrofuran can undergo auto-oxidation and form peroxides (1), and for this reason tetrahydrofuran is commonly sold with stabilizers or free-radical inhibitors to scavenge peroxides that form during storage. These inhibitors cause erroneous peaks during PTH-AA separation, so PE/ABD does not add them to their HPLC solvents (personal communication with PE/ABD technical support). Because every pmol matters in high sensitivity protein sequencing, a method for monitoring peroxide levels in solvent A is needed.
Peroxide levels can be quickly and easily measured with EM Quant Peroxide test strips. The strips are 0.6 cm x 7.5 cm long with an indicator area of 0.6 cm2 and a working pH range of 2-12. According to the manufacturer, the strips contain peroxidase, which transfers oxygen from peroxides to an organic redox indicator, producing a blue-colored oxidation product on the strip. A standard strip with a color scale is provided for measurements of 0, 0.5, 2, 5, 10, and 25 mg H2O2 per liter (ppm). The test is very simple and fast but more qualitative then quantitative. For aqueous solutions, solvent A is spotted on the indicator zone of the test strip and compared to the standard strip's color scale after 15 secs. Organic solvents can be tested by dipping the test strip in solvent for 1 sec, waving it in the air for 3-30 secs to evaporate excess solvent, dipping it into distilled water for 1 sec, gently shaking off excess water, and then comparing to the standard strip after 15 secs. The peroxide test strips are from EM Science and can be purchased in the U.S. from VWR Scientific (catalogue number EM 10011 1, pack of 100 strips for $28.20) and in Europe from E. Merck (Postfach 41 19, D 64271 Darmstadt, Federal Republic of Germany, Tel: (0 61 51) 720, Telex: 4 19 328 0 em d).
I began using this method for measuring peroxide contamination because I observed decreased yields of PTH-Lys when using some prepackaged lots of PE/ABD solvent A (5% THF) and more often when using homemade solvent A. Because this was from a past observation with no systematic data to support my conclusions, I decided to document the effects of peroxides on PTH-AA's and correlate these effects to the amounts of peroxides detectable with test strips. Fresh bottles of Solvent A3 (3.5% THF) were first checked for peroxides. A lot from 1993 tested at 0.5 mg per liter (ppm) whereas a fresh bottle of a 1995 lot contained no detectable peroxides. PE/ABD Premix ion-pairing reagent (20 ml) was added to the pre-tested 1995 bottles, and these were used for the experiments described below.
First I used hydrogen peroxide as the peroxide source to see how much PTH-Lys was effected. Hydrogen peroxide (10 ml of 30% w/w from Sigma) was added to 100 ml of PE/ABD solvent A3 (3.5% THF). The peroxide test was performed, and the peroxide level was observed to be greater than 25 mg per liter (ppm). From this stock further dilutions were made to a total volume of 200 ml each. The dilutions, the calculated volumes of H2O2 per liter of solvent A, and the peroxide levels as measured with EM Quant peroxide test strips were as follows: 1:2 dilution, 50 ml H2O2 per liter, 25 mg per liter; 1:4 dilution, 25 ml H2O2 per liter, 10 mg per liter; 1:10 dilution, 10 ml H2O2 per liter, 5-10 mg per liter; and 1:100 dilution, 1 ml H2O2 per liter, 0.5-2 mg per liter. Each solvent mix was used with a PE/ABD 476A sequencer configured with a new PE/ABD HPLC PTH-column and an 80 ml injection loop. The HPLC pumps were purged twice between solvent changes before collecting data for one blank run and three 40 pmol PTH-AA standard chromatograms with the PE/ABD model 610 data analysis system. Peak areas were used for quantitating each PTH-AA. The three standard runs with each peroxide-spiked solvent were averaged and tabulated. Figure 1 shows the percent area loss for each PTH-AA compared to a 0 mg H2O2 per liter solvent A control.
As seen in Figure 1, all primary PTH-AA's were reduced by at least 5%, and PTH-Met showed a 75% loss at a peroxide concentration of 25 mg per liter. Because this was a more dramatic effect than I had observed in the past, I decided to repeat the experiment with aging Aldrich THF (catalogue number 27,038-5) that contained peroxide levels above the detectable range of the indicator strips (indicator turned black). THF (1.5 ml) was added to 100 ml of fresh solvent A3, giving a final THF concentration of 5% and a peroxide level of 10-25 mg per liter. The dilutions and the peroxide levels as measured with test strips were as follows: no dilution, 10 mg per liter; 1:4 dilution, 2 mg per liter; 1:10 dilution, 0.5-2.0 mg per liter; and no additional peroxide. Standard chromatograms were obtained following the protocol described above, and the results are shown in Figure 2. As in the previous experiment with hydrogen peroxide, all PTH-AA's were affected but to a lesser degree, except for PTH-Lys. From past recollections, noticeable losses of PTH-Lys were seen when homemade solvents contained peroxides at levels of 2.0 - 5.0 mg per liter, using Aldrich THF stored under nitrogen at 4šC. Based on these current observations, a 20% decrease in PTH-Lys yield can be expected when peroxides reach the 2.0 mg per liter level.
Lastly, I examined the effects of peroxide contamination on high sensitivity sequencing. I diluted the standard so that 4 pmol of each PTH-AA standard was injected on the column and examined peroxide effects as described above. Aldrich THF (1.0 ml) was added to 100 ml of solvent A3, and test strips showed the peroxide level of this stock was 10 mg per liter. The dilutions and the peroxide levels as measured with test strips were as follows: no dilution, 10 mg per liter; 1:2 dilution, 5 mg per liter; 1:4 dilution, 2 mg per liter; and no additional peroxide. The results are shown in Figure 3. At high sensitivity there is an overall increase in percent loss of all PTH-AA's. However, at very high peroxide levels, the PTH-AA's are affected less than at the lower peroxide levels, contrary to expectations based on the data shown in Figures 1 and 2. The reason for this is unclear.
This study shows that PTH-Lys tends to be the PTH-AA most sensitive to oxidation, as originally observed and best illustrated in Figure 2. Hydrogen peroxide or organic peroxides affect PTH-AA yields in characteristically different ways. Hydrogen peroxide is much more reactive toward most PTH-AA's than the peroxides found in the HPLC solvents.
Storing THF under a blanket of argon or nitrogen gas is thought to help prevent peroxide formation. Several other suggestions were made during the electronic mailing list discussion. Mike Rohde (Amgen) suggested diluting the entire 1 liter bottle of THF into 20 liters of Burdick and Jackson water and storing. His group had fewer problems with peroxide formation in diluted THF stocks. Carol Beach (Univ. of Kentucky) suggested storing 1 liter bottles of THF in a -20šC freezer, removing the bottles just long enough to pour out the required amount, and then re-blanketing them with inert gas. A final point to consider is the presence of DTT in R4A (25% TFA in H20) and how it can affect the level of PTH-AA oxidation. This past February, PE/ABD sent a letter to its protein sequencer users notifying them that many bottles of R4A gave decreased PTH-Lys recovery in standards and samples. They attributed this PTH-Lys destruction to an unidentified oxidizing contaminant in R4A, which also produced a large oxidized DTT peak in the early part of the chromatogram. If low PTH-Lys yields are observed and detectable peroxides are not present in solvent A, solvent R4A could be the source of the problem.
References
The author may be contacted at Novo Nordisk Biotech, Inc., 1445 Drew Ave., Davis, CA 95616.
Return to the The ABRF Home Page